Course_June22 26_2026 - I2PC/scipion-courses GitHub Wiki

Model Building

Model Building Agenda

June 24th

• 14:00 - 14:45: General introduction to model building in cryo-EM (M)
  • Example 1: Naive model with Coot
• 14:45 - 15:15: Introduction to the data used in the practical sessions (human_haemoglobin) (details here) (M)
  • Review Kahoot students, teacher (R 000)
• 15:15:16:00
  • Homology modeling workflow (R)
  • Import map and sequence. Input data (R)
  • Map preprocessing (M)

TAKE HOME MESSAGE: Try to improve the visual aspect of your map in terms of connectivity and/or details before tracing it.

• 16:00 - 16:30: coffee break
• 16:30 - 17:30: Getting a first estimate of the model "de novo" (AlphaFold & ModelAngelo) and by sequence homology. (R-R-M)

  TAKE HOME MESSAGE: Use the structure of a good homologous sequence if you have it, otherwise try ModelAngelo if your map's resolution is good enough, alternatively try to predict the structure with AlphaFold. Remember to double check the model with the experimental data.

• 17:30: End of the day

June 25th

• 09:30-09:45: Review yesterday's work (R)
• 09:45-10:30: Structure analysis: Structure comparison (M)
• 10:30-11:00: Rigid Fitting of Initial Models (R)
• 11:00-11:30: Coffee break
• 11:30-13:00: Flexible fitting (Coot) and Validation
  • 11:30-12:15 Coot Tutorial (R)
  • 12:15-13:00: Flexible fitting (Phenix & Refmac) and Validation (M-M)
• 13:00-14:00: Lunch Break
  • 14:00-15:00: Flexible fitting (Phenix & Refmac) and Validation (M-M) (cont)
  • 15:00-15:30: Review Kahoot students, teacher (R: 0001 - revision)
• 15:30-16:00: Model building of human Hgb beta subunit (R-M)
• 16:00-16:30: Coffee break
• 16:30-17:00: Model building of human Hgb beta subunit (R-M) (cont)
• 17:00-17:30 Building the model of the whole protein
  • Chain merge & symmetrization (R)
• 17:30:End of the day

  TAKE HOME MESSAGE: Models are always wrong... but some are useful.

June 26th

• 09:30-09:45: Review yesterday's work (R)
• 9:45-10:00: Submission of structures to EMDB. Understanding the validation report. Haemoglobin PDB full validation report (R)
  • Comparison validation scores PDB structure(M)
• 10:00-11:00: Structure analysis: chain contacts Structural search with DALI Structural search with FOLDSEEK (M)
• 11:00-11:30: Coffee break
• 11:30-13:00: AI new developments (M)
  • Introduction to protein design
  • Example with CARBonAra
• 13:00-13:00: Course Questionnaire (R)
• 13:00-14:00: Lunch
• 14:00: End of the Course
• XX:XX: * Review Kahoot students, teacher (R: Review Model Building)

---

Course Material

• Presentations

  • Introduction to Model Building
  • Introduction to AlphaFold2
  • Introduction to ModelAngelo
  • Scipion model building tutorials:
    1. Build initial model by Homology
    2. Build initial model with AlphaFold
    3. Build initial model with ModelAngelo
  • Specimens:
    1. Human haemoglobin
    2. Human ion channel-like protein TACAN

• Practical working sessions:

  1. Homology modeling workflow
  2. AlphaFold & ModelAngelo modeling workflow

Data resources and IDs

• How to access to the remote desktop

• BlastP (search for homologues)

• Human haemoglobin

  • 3D map: emd_3488.map, available directory Data
  • sequence:
    • alpha subunit: UniProt ID P69905 (Homologous sequences: P01958, P01948, P81023, P83134)
    • beta subunit: UniProt ID P68871 (Homologous sequences: P02062, P02057, G1U9Q8, P83133)
  • PDB reference: 1pbx (antartic fish: Pagothenia bernacchii)

• Human ion channel-like protein TACAN

  • 3D map: emd-31482.map, available directory Data
  • sequence:
    • UniProt ID Q9BXJ8

• Coot Tutorial

  • 3D map
  • atomic model

• ChimeraX Tutorial

• EMDB submission link

  • Haemoglobin in PDB: ID 5ni1
    • Haemoglobin PDB full validation report

• Contacts AD type V (Zhou)